QW

Qwen3-4B-Instruct-2507

QwenGGUF by unslothUpdated 2025-08-20
Download on Hugging Face
Parameters4B
Context256K
Layers36
Architectureqwen3
Smallest quant1.1 GB
Quick answer

Qwen3-4B-Instruct-2507 needs about 4.2 GB of memory at Q4_K_M (a 2.5 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 26 quants, from 1.1 GB (IQ1_S) to 8.1 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.8 GB.

Qwen3-4B-Instruct-2507 – 4 B parameter instruction model with 256 K context

Qwen3-4B-Instruct-2507 is a 4 B parameter causal language model from the Qwen family. It supports a native context length of 262,144 tokens and is optimized for instruction following, reasoning, coding and tool usage. The model is released under the Apache‑2.0 license.

Good at

  • Instruction following
  • Long‑tail knowledge across languages
  • Tool‑calling and agentic tasks

Limits

  • No thinking mode support
  • Requires 8 k context for memory estimates
  • No vision capability

Best for

  • Chatbots needing long context
  • Code generation with tool calls
  • Research prototyping with small VRAM

Who it’s for. Developers and researchers who need a compact instruction model with extended context. The model is suitable for local deployment on systems with limited GPU memory.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S1.1 GB 2.8 GB
Fits Get file →
IQ1_M1.1 GB 2.9 GB
Fits Get file →
IQ2_XXS1.3 GB 3.0 GB
Fits Get file →
IQ2_M1.5 GB 3.2 GB
Fits Get file →
Q2_K1.7 GB 3.4 GB
Fits Get file →
Q2_K_L1.7 GB 3.4 GB
Fits Get file →
IQ3_XXS1.7 GB 3.4 GB
Fits Get file →
Q2_K_XL1.7 GB 3.4 GB
Fits Get file →
Q3_K_S1.9 GB 3.6 GB
Fits Get file →
Q3_K_M2.1 GB 3.8 GB
Fits Get file →
Q3_K_XL2.1 GB 3.8 GB
Fits Get file →
IQ4_XS2.3 GB 4.0 GB
Fits Get file →
Q4_02.4 GB 4.1 GB
Fits Get file →
IQ4_NL2.4 GB 4.1 GB
Fits Get file →
Q4_K_S2.4 GB 4.1 GB
Fits Get file →
Q4_K_M2.5 GB 4.2 GB
Fits Get file →
Q4_K_XL2.5 GB 4.3 GB
Fits Get file →
Q4_12.6 GB 4.3 GB
Fits Get file →
Q5_K_S2.8 GB 4.5 GB
Fits Get file →
Q5_K_M2.9 GB 4.6 GB
Fits Get file →
Q5_K_XL2.9 GB 4.6 GB
Fits Get file →
Q6_K3.3 GB 5.0 GB
Fits Get file →
Q6_K_XL3.7 GB 5.4 GB
Fits Get file →
Q8_04.3 GB 6.0 GB
Fits Get file →
Q8_K_XL5.1 GB 6.8 GB
Fits Get file →
F168.1 GB 9.8 GB
Too big Get file →

For 8 GB VRAM, use Q4_K_M (2.5 GB file, ~4.2 GB VRAM at 8 k context). For 12 GB VRAM, Q5_K_M (2.89 GB file, ~4.6 GB VRAM) or Q6_K (3.31 GB file, ~5 GB VRAM). Larger VRAM allows Q8_0 (4.28 GB file, ~6 GB VRAM).

What GPU do you need for Qwen3-4B-Instruct-2507?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is IQ4_XS (2.3 GB file, about 4.0 GB in use).
  • 6 GB VRAM: the best quant that fits is Q8_0 (4.3 GB file, about 6.0 GB in use).
  • 8 GB VRAM: the best quant that fits is Q8_K_XL (5.1 GB file, about 6.8 GB in use).
  • 12 GB VRAM: the best quant that fits is F16 (8.1 GB file, about 9.8 GB in use).

Which Qwen3-4B-Instruct-2507 quant should you choose?

Q4_K_M is the usual starting point for Qwen3-4B-Instruct-2507: it balances file size and quality well. Here it is a 2.5 GB file.

The Q8_0 file is 4.3 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is IQ1_S at 1.1 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen3
Parameters4B
Layers36
Trained context262,144 tokens
KV cache (fp16)144 KB per token

Source

Quantized byunsloth
FamilyQwen
Licenseapache-2.0
Downloads (HF)449,529
Last updated2025-08-20

Frequently asked questions about Qwen3-4B-Instruct-2507

How much VRAM does Qwen3-4B-Instruct-2507 need?

About 4.2 GB at Q4_K_M with an 8K context window. That is the 2.5 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Qwen3-4B-Instruct-2507 on an 8 GB GPU?

Yes, with Q8_K_XL: the 5.1 GB file needs about 6.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Qwen3-4B-Instruct-2507 GGUF?

It depends on the quant: IQ1_S is 1.1 GB, IQ1_M is 1.1 GB, IQ2_XXS is 1.3 GB, IQ2_M is 1.5 GB, Q2_K is 1.7 GB, Q2_K_L is 1.7 GB and 20 more.

What context length does Qwen3-4B-Instruct-2507 support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.2 GB at 8K and 4.8 GB at 32K.

Where can I download Qwen3-4B-Instruct-2507 GGUF?

The files are on Hugging Face in the unsloth/Qwen3-4B-Instruct-2507-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Qwen3-4B-Instruct-2507 use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Qwen3-4B-Instruct-2507 on 8 GB VRAM?

Yes, a Q4_K_M quantized model (~2.5 GB file) will use about 4.2 GB VRAM at 8 k context, leaving headroom for the host process.

Does the model support vision or image input?

No, Qwen3-4B-Instruct-2507 is a text‑only causal language model and does not process images.

What is the maximum context length I can use?

The model natively supports 262,144 tokens, but memory usage increases with context length; reduce to 32,768 if OOM occurs.

Other GGUFs of Qwen3-4B-Instruct-2507

Similar models

Browse all →

More from unsloth

All by unsloth →

More Qwen models

All Qwen →

Other small models

All small models →